Maximizing Efficiency: Harnessing second-hand computer systems for parallel computing in university micro cloud
Georgi Kostadinov Petrov, Rosen I. Pasarelski, David Yavorov, Mile Milanoski · Godišnik Telekomunikacii/Godišnik na Departament Telekomunikacii · 2024
This article explores the utilization of second-hand computer systems in the context of parallel computing for physics simulations at New Bulgarian University. With the increasing demand for computational power in scientific research, particularly in the field of physics, the affordability and accessibility of high-performance computing (HPC) resources become crucial. However, acquiring brand-new HPC systems can be cost-prohibitive for many laboratories and institutions. In this paper, we investigate the feasibility and efficacy of repurposing second-hand computer hardware to construct parallel computing clusters tailored for physics simulations. We discuss strategies for selecting, refurbishing, and integrating these systems into a parallel computing environment, highlighting the potential benefits and challenges. Through case studies and performance evaluations, we demonstrate the practicality of this approach in achieving significant computational capabilities at a fraction of the cost of new systems. Our findings underscore the importance of resourcefulness and sustainability in advancing computational physics research, paving the way for cost-effective solutions to address complex scientific challenges.